Lateblight Advisor: A Weather-Driven Decision Support Dashboard for Potato Late Blight Risk and Fungicide Scheduling

This dataset describes Lateblight Advisor, an interactive Python Shiny dashboard for potato late blight decision support. The application lets users configure a field scenario (dates, country, and variety/resistance class), select a field location on an interactive map, retrieve historical and forecast weather from WeatherAPI, and compute weather-driven risk metrics including Blight Units (BU) and Fungicide Units (FU). Model outputs are summarized through metrics, plots, a calendar view, and a recommendations table to support fungicide application timing. The application also includes an optional LLM-assisted retrieval/search component using OpenAI and Tavily, plus a local Chroma vector store for retrieval-augmented responses. The full source code is publicly available on GitHub and is referenced from this dataset record.